UK startup Quinas looks to solve memory problem Intel, others have not
It’s a familiar promise: memory capable of retaining its content like flash memory, but with the speed of DRAM. Intel and a slew of startups have tried and failed at the product, but a UK startup may be the one to deliver it.
Quinas Technology, a British semiconductor company, an IQE a supplier of compound semiconductor wafer products, has announced the successful completion of a joint project to develop a scalable production method for the world’s first quantum-powered universal memory called ULTRARAM.
The project saw IQE successfully scale up the manufacture of compound semiconductor layers initially developed by Lancaster University to an industrial process, the first step towards the commercial production of packaged ULTRARAM chips.
ULTRARAM is a dual-use technology invented at Lancaster. It combines the data storage capability found in flash memory with the speed and endurance of DRAM, while providing significantly improved energy efficiency with a considerably lower power draw.
Quinas also makes some astonishing claims on ULTRARAM’s durability. NAND flash memory can only withstand around 10,000 program/erase cycles. After that, the cells start to degrade and lose the ability to hold memory contents. In contrast, ULTRARAM has demonstrated degradation free operation more than 10 million program/erase cycles. This is due to the low-voltage, ultra-low-energy program/erase process enabled by quantum resonant tunnelling, the company stated.
“They’re in the pretty early stages of developing a technology. It promises to maintain bits longer than most, but there are lots of practical issues to work out,” says Jim Handy, president of Objective Research, which specializes in the memory market.
For starters, the memory is built on what is called a” III-V technology,” a class of semiconductor materials that are composed of elements from groups III and V of the periodic table, the company stated. These materials have unique properties that make them ideal for use in electronic devices such as transistors, LEDs, and solar cells.
Any technology that’s not pure silicon causes trouble when built on a standard silicon CMOS logic process. This has been one of the leading reasons why other attempts at high-speed non-volatile memory technologies like MRAM, ReRAM, PCM, and FRAM haven’t gotten off the ground, Handy said.
Handy notes the memory’s voltage of 2.6 volts Is orders of magnitude lower than the 20 volts required by NAND flash memory. That should translate into much lower electric bills and much longer life for the storage systems.
The news of successful manufacturing is an important step because it was the economies of scale that did in Optane, Handy said. “The wafer volume never ramped high enough to drive down the costs. There’s a chicken-and-egg problem that confronts any new technology as it vies to replace an established market leader, which Optane tried to do with DRAM. The volume needs to get high enough to drive down the cost, and the price must be low enough to grow the market volume. Intel tried to subsidize the ramp long enough to achieve this, but gave up after losing about $10 billion,” he said.
